Borrowing it
Nothing to install: this file belongs to nimbus-agent/nimbus-mcp. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/nimbus-agent/nimbus-mcp/main/.claude/commands/nimbus-mcp-boundaries.mdgit clone --depth 1 https://github.com/nimbus-agent/nimbus-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/nimbus-agent/nimbus-mcp/nimbus-mcp-boundaries)<a href="https://agentmods.dev/commands/nimbus-agent/nimbus-mcp/nimbus-mcp-boundaries"><img src="https://agentmods.dev/badge/commands/nimbus-agent/nimbus-mcp/nimbus-mcp-boundaries/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/nimbus-agent/nimbus-mcp/nimbus-mcp-boundaries"><img src="https://agentmods.dev/badge/commands/nimbus-agent/nimbus-mcp/nimbus-mcp-boundaries.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00066 | $0.02293 |
| Opus 5 | $0.00033 | $0.01146 |
| Sonnet 5 | $0.00013 | $0.00459 |
| Haiku 4.5 | $0.00007 | $0.00229 |
Grade A, and why
nimbus-mcp-boundaries scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 144 lines — stays where its author put it; the contents beside it link to each section on GitHub.
nimbus-mcp — boundaries, guards, and what the gates miss
CLAUDE.md and CONTRIBUTING.md already carry the rules: zero runtime dependencies,
MIT and never importing from the AGPL monorepo, resolution order NIMBUS_BIN → PATH
→ known install dirs. Those two files win on doctrine. This one is the part that is
expensive to rediscover: which guard is load-bearing, which is theatre, and what a
change here has to touch in the other repo.
1. The installer contract is two-sided, and this side is the weak half
src/installer-contract.ts exports two literals vendored from the monorepo's
scripts/install/lib/paths.ts (resolveInstallDir). installer-contract.test.ts
asserts CANDIDATE_DIRS' first entry per platform is built from them.
That test cannot fail when the installer changes. It compares this repo against its
own copy. If someone changes resolveInstallDir in the monorepo tomorrow, everything
here stays green and every PATH-less MCP client starts reporting "Could not find the
Nimbus CLI" against a working install.
The half that actually catches it is
scripts/structure-audit/check-launcher-installer-contract.ts in the monorepo, run by:
install-smoke.yml— PR-time, triggered onscripts/install/**. Catches their change.org-drift-sweep.yml— scheduled, clones this repo. The only thing that catches a change made here.
Practical consequences:
- That script parses the constant names with a regex, and the literal forms
(
String.rawbacktick for win32, double-quoted for posix). Renaming a constant or switching quote style breaks the parser — and a broken parser reports a mismatch, which reads like real drift. Change both sides in one sitting. - Editing a value here without a matching monorepo change will not fail this repo's CI. It fails the sweep, later, in another repository.
2. src/index.ts cannot be unit-tested, and that is deliberate
It is a shebang bin entry with top-level side effects: it resolves, spawns a child, and
wires exit. Importing it from a test executes it — spawning a process and
potentially calling process.exit. That is why it carries sonar.coverage.exclusions
and why the monorepo exempted the same file while it lived there.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 144 lines · 66 tokens per session scan A f2720d9f857c
nimbus-mcp-boundaries is a command published in the GitHub repository nimbus-agent/nimbus-mcp (0 stars, last pushed yesterday), licensed MIT. It adds 66 tokens to every session and 2,293 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
checklist
Generate a custom checklist for the current feature based on user requirements.
clarify
Identify underspecified areas in the current feature spec by asking up to 5 highly targeted clarification questions and encoding answers back into the spec.
specify
Create or update the feature specification from a natural language feature description.
analyze
Perform a non-destructive cross-artifact consistency and quality analysis across spec.md, plan.md, and tasks.md after task generation.
converge
Assess the current codebase against the feature's spec, plan, and tasks, then append any remaining unbuilt work as new tasks to tasks.md so implement can complete it.
implement
Execute the implementation plan by processing and executing all tasks defined in tasks.md.